作者位次 #2/7ACL 2026Findings· Findings of ACL
X -Router: Decoupling Knowledge and Reasoning for Cost-Effective LLM Inference
Zixuan Wang · Yinze Ding · Zihan Wang · Jinyu Guo · Zhenhong Zhou · Junhao Dong · Chaomeng chen
摘要
Large Language Models (LLMs) are often augmented with Retrieval-Augmented Generation (RAG) and Chain-of-Thought (CoT) prompting, yet static “always-on” use is computationally wasteful. Existing adaptive methods typically optimize a single axis, overlooking that evidence need and reasoning depth a...